A formal approach to modeling and analyzing human taskload in simulated air traffic scenarios

摘要

In complex systems, like the modern air traffic system, human operator taskload can have a profound influence on how well the system performs. Because of the system's complexity, however, it can be difficult to determine all of the situations where taskload issues can arise. Simulation and formal verification have been used separately to explore human taskload in complex systems. However, both have problems that limit their usefulness. In this paper, we describe a formal modeling architecture designed to enable the discovery of interesting human operator taskload conditions though the synergistic use of formal verification and simulation. This architecture formally represents original simulation constructs using computationally efficient abstractions that ensure that temporal and ordinal relationships between simulation events (actions) are represented realistically. Taskload for each agent is represented based on a priority queue model where only a limited number of actions can be performed or remembered by a human at any given time. We provide an overview of this architecture, discuss its essential features, and describe the mathematical foundations needed for its instantiation. We present insights into its capabilities for finding interesting taskload conditions by formulating several checkable specification properties. The implications of this architecture are discussed in terms of its broader supported analysis method and directions for future work are explored.